This article first appeared in MIT Sloan Management Review on October 7, 2020.
Companies can deliver critical knowledge services more efficiently by embracing distributed computing’s modular approach to allocating resources.
When companies shifted to remote working in response to the coronavirus pandemic, many discovered the hard way that decades of offshoring critical knowledge services around the globe to maximize efficiency left them vulnerable to abrupt disturbances to business continuity. Suddenly, mortgages were not approved. Customer support hotlines were unavailable. Paychecks were delayed.
Disruptions — major and minor — will continue to impact businesses even after COVID-19 outbreaks recede. Other risks, such as cyberattacks, trade wars, and extreme weather events caused by climate change, linger on the horizon.
To better prepare their services for the future, companies can take a page from distributed computing principles, which have defined points to receive commands and return results but then dynamically rebalance tasks across computational nodes to maximize efficiency. By adopting a similar “any-to-any” approach to services, companies can improve their performance and resilience by breaking up services into modules that can be reallocated across multiple networks and locations based on availability rather than geography.
Companies first started shifting the processing of services offshore at scale during the internet boom in the 1990s as a way to achieve new back-office efficiencies. But since the turn of this century, the process offshoring business has transformed from one centered around low labor costs in Asia into a $200 billion industry employing tens of millions of specialists in everything from customer service to research and development around the world.
Today, when a bank issues a mortgage in the United Kingdom, it might originate in the U.K. but be processed in Ireland, paid out from India, and serviced again in any of these locations. But during a global crisis, this entire process fails if any of these steps is disrupted.
Across companies and sectors, technology teams follow any-to-any principles that allow coders to collaborate across platforms, tools, and locations. These same principles can allow companies to create parallel service processing environments that permit service modules to be run and shifted flexibly among teams and locations. As a result, if a customer’s request cannot be addressed in Ireland, the task can be delegated to a team in the Philippines or to another team on standby in another part of the world.
While delivering any-to-any services may sound straightforward, it requires companies to shift their processing to multiple, alternate locations, when most of them currently have service processing operations parked in data centers in single offshore locations.
While delivering any-to-any services may sound straightforward, it requires companies to shift their processing to multiple, alternate locations, when most of them currently have service processing operations parked in data centers in single offshore locations. Teams operating these services then need to work collaboratively with people in many more countries — and possibly even in different companies — across standardized technologies and forms of data.
That means “live” processing capabilities have to be replicated across more teams and locations. New systems need to be fully tested for use across locations as well as multiple teams’ potentially less secure remote work environments. One leading Indian services company, for example, aims to have at most 25% of its workforce working on premises while offering virtual access to all of its systems and data to people working remotely. In addition, people need to be retrained in more than one type of activity — such as loan servicing — to shift resources more nimbly to where there is greater demand. Individuals should then be deployed to these alternate activities periodically so that they can retain the expertise, enabling the organization to redeploy tasks in evolving circumstances.
So, what steps should companies take now to ensure that they are ready for the service economy’s next any-to-any phase?
1. Determine your critical processes. First, identify and focus on the processes that are critical for your company to operate and deliver its core obligations over different time frames. For example, payment processing in a bank is critical immediately, whereas regulatory reporting becomes critical over weeks or months. Then examine how well these different processes were performed during periods of disruption.
2. Identify single points of failure. Next, analyze where single points of failure could cripple critical processing systems. Scour locations, people, systems, data, tools, infrastructure, and any third parties involved for potential vulnerabilities. Test how these choke points may migrate across systems and processes depending on the scope, duration, and severity of a crisis — be it a pandemic, cyberattack, or natural disaster. That way, you can determine which challenges your any-to-any setup should solve for and where redundancy is required.
3. Design your any-to-any network. Armed with this information, reshape your organization into an any-to-any network. Define the minimum operational performance for critical processes. For example, a bank may decide that, at a minimum, 60% of mortgage applications must be processed within five days. That means the bank will have to set up an any-to-any model that can shift 60% of the bank’s mortgage capacity among teams in different locations, subject to jurisdictional legal and regulatory restrictions, within five days if any point along the process fails.
This requires breaking up each of these processes into a series of interconnected modules that optimize resilience and recoverability through any-to-any connections, with a clearly identified plan for differing process options and associated controls. For example, one organization we worked with made it possible for customers to sign up for a new service in half as much time and with fewer interruptions by dividing its customer onboarding process into multiple required steps that could be processed in parallel through multiple, redundant systems.
4. Upgrade workflow and communication tools. Establish common tools (and fallbacks) across your organization so that workflows can be managed seamlessly across locations and teams. Most companies have recovery plans linked to physical sites that require on-premises access to work with systems and data. But in an any-to-any service model, workflows should be managed and accessed virtually, with appropriate controls and access rights globally shared across operationally involved teams. That includes monitoring and ensuring the operational effectiveness of the processes, controls, tools, assets, and modular capabilities.
5. Establish any-to-any operations. To fit daily operations to a distributed any-to-any model, companies will need an agile governance model, whereby process owners in each team are responsible for managing their process modules — including decision-making on whether they can take on work delegated from other teams or need to shift work in times of crisis. These can be facilitated with centralized virtual “marketplaces” in which service models can be reallocated dynamically among teams based on demand and supply for processing. Simultaneously, senior managers should oversee the portfolio of activities underway to ensure the strategic alignment and governance of this distributed decision-making.
Building Efficiency and Resilience
Offshored knowledge services have enabled companies to reach many more customers than previously thought possible around the clock at an unprecedented speed. But disruptions caused by the current pandemic have shown that many of these global service processing systems are too rigid to function in times of crisis.
In response, some companies are starting to bring more service processing in-house or onshore. But the companies that resist this temptation and develop much more resilient any-to-any global service processing operations will reap the benefits of offering services to more customers in even more locations. In the process, they will redefine the rules for a new “right-shoring” phase of the service economy.